The ser_par demo illustrates the tradeoffs between serial 
and parallel implementation of a circuit.

Demo1: File sh_leftright.mcl

For sh_leftright.mcl, the parameters are
        w               data width
        arch    architecture select, arch=0 selects parallel arch.
                                     arch=1 selects serial arch.
 
Pull down the "File" menu and select "Get Parameters".
This will extract the default parameters from the
module.  For demo purposes the default parameter
values can be used as is. Click "Do All".
Now, change the parameter "arch" from 
0 to 1 and click "Do All".
 
Pull down the "View" menu and select "Table Summary".
Note the area and timing #'s.
As expected the serial architecture is smaller but slower than the 
parallel one.


Demo2: File var_coeff.mcl and const_coeff.mcl
This demo illustrates the trade-off of area vs delay related to serial 
and parallel implementations of a color space converter circuit which 
is commonly used in video front-end processors and other video and 
graphic applications.
 
There are 2 modules related to this circuit one for the color space 
converter using constant co-efficents (const_coeff.mcl) and the other 
for the variable co-efficients (var_coeff.mcl). For each of these, 
we have serial and parallel architecture implementations.
 
The parameters used in the modules are
        arch    architecture select, arch=0 selects parallel arch.
                                     arch=1 selects serial arch.
        fa      final adder architecture fastcla, clsa, cla, csa, ripple.
In case of the color space converter with constant co-efficients 
there are 9 other parameters which are essentially the constant 
co-efficients required for the color conversion.
 
Start with the color space converter with variable co-efficients by 
entering the filename var_coeff.mcl in the Input File field of the GUI.
Build circuits for both the architectures (serial as well as parallel) and
compare the area/delay between them. Then build circuits for the same
converter when the co-efficients are constant by using the const_coeff.mcl 
file and see the impact on area/delay for each architecture and compare 
it with the previous results.
